Effect of particle size distribution of raw powders on pore size distribution and bending strength of Al2O3 microfiltration membrane supports

+Model JECS-9695; No. of Pages 7

ARTICLE IN PRESS Available online at www.sciencedirect.com

ScienceDirect Journal of the European Ceramic Society xxx (2014) xxx–xxx

Effect of particle size distribution of raw powders on pore size distribution and bending strength of Al2O3 microfiltration membrane supports Qibing Chang a,∗ , Yulong Yang a , Xiaozhen Zhang b , Yongqing Wang a , Jian-er Zhou a , Xia Wang a , Sophie Cerneaux b , Li Zhu c,d , Yingchao Dong c,d a

School of Materials Science and Engineering, Jingdezhen Ceramic Institute, Xianghu, Jingdezhen, Jiangxi Province 333403, PR China b Institut Europeen des Membranes UMR 5635, Place Eugene Bataillon, 34095 Montpellier Cedex 5, France c Institute of Urban Environment, Chinese Academy of Sciences, PR China d Ningbo Urban Environment Observation and Research Station-NUEORS, Chinese Academy of Sciences, PR China Received 14 March 2014; received in revised form 11 May 2014; accepted 2 June 2014

Abstract To lower the sintering temperature of Al2 O3 microfiltration membrane support, Al2 O3 powders with particle size distribution of tri-modal are chosen. The results show that the function of fine Al2 O3 grains depends on their agglomeration state: if fine Al2 O3 grains distribute discretely, the bending strength of the support increases along with a slight increase in porosity; however, the aggregated fine grains are harmful to both bending strength and pore size distribution of the support. The bridging of medium Al2 O3 grains between coarse grains contributes to increase the bending strength, but has less effect on porosity. The addition of medium (and/or fine) Al2 O3 powder has less effect on the pore size distribution of the support if only coarse Al2 O3 grain forms the support’s framework, which suggests a new way to prepare the support with both high bending strength and high porosity at low temperature. © 2014 Elsevier Ltd. All rights reserved. Keywords: Porous ceramic; Alumina membrane support; Particle size distribution; Pore size distribution; Bending strength

1. Introduction Ceramic membranes have been gaining more attentions in industrial application due to their excellent thermal, chemical and mechanical stability.1–4 Generally, ceramic membranes have an asymmetric structure consisting of a porous thin layer with separation function on a support. An excellent microfiltration membrane support should possess high mechanical strength, high permeability, and narrow pore size distribution, which mainly depend on the raw materials, sintering temperature and fabrication methods.5 To achieve the high permeability, coarse alumina powder is generally chosen for the commercialized microfiltration membrane support. However, the poor sinteringactivity of coarse alumina grains results in high cost of this kind support because high sintering temperature is required to obtain

∗

Corresponding author. Tel.: +86 798 8499162; fax: +86 798 8494973. E-mail address: [email protected] (Q. Chang).

sufficient bending strength.6 Though some raw materials with low melting temperature had been used to prepare the ceramic support,6–8 alumina support is still popular due to its good bending strength and excellent acid/alkali resistance.8 Some sintering aids, such as TiO2 ,4,7 boehmite,9 kaolin,10 and clay,11 were added to decrease the sintering temperature. However, these materials weaken the bending strength and acid/alkali-resistance of the support in some degree.11 Ceramic microfiltration support is a kind of porous ceramic, whose bending strength is proportional to the neck areas among the grains. The particle size has an important effect on the neck area among the grains through sintering. To a pair of grains with different particle sizes, it has been proved that fine grains disappear and coarse grains coarsen through surface diffusion or grain boundary diffusion.12–16 To an array of coarse-fine-coarse grains, the neck area among coarse Al2 O3 grains increases due to the immigration of fine grains. Taking this rule into account, the support is designed using alumina powders with the particle size distribution of tri-modal as raw materials. The support sintered at

http://dx.doi.org/10.1016/j.jeurceramsoc.2014.06.001 0955-2219/© 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Chang Q, et al. Effect of particle size distribution of raw powders on pore size distribution and bending strength of Al2 O3 microfiltration membrane supports. J Eur Ceram Soc (2014), http://dx.doi.org/10.1016/j.jeurceramsoc.2014.06.001

+Model JECS-9695; No. of Pages 7 2

ARTICLE IN PRESS Q. Chang et al. / Journal of the European Ceramic Society xxx (2014) xxx–xxx Table 1 Compositions of W50, W10 and W1 used for preparing supports. Powder

W50/wt%

W10/wt%

W1/wt%

80

16 10 6

4 10 14

S4 S5 S6

70

24 15 9

6 15 21

S7 S8 S9

60

32 20 12

8 20 28

Sample label S1 S2 S3

Fig. 1. Particle size distributions of W50, W10 and W1.

a relative low sintering temperature can own both high bending strength and high porosity. Simultaneously, the broader particle size distribution contributes to the sintering because of the high green compact density.12,17–19 In another words, the sintering of the porous ceramic can be realized by the sintering-induced immigration of the fine grains at a relatively lower temperature without sintering of coarse grains. This method gives a promising way to obtain ceramic membrane support with high porosity without degrading bending strength. However, the broad particle size distribution may also lead to the difference of localized densification,17 which results in a broad pore size distribution.20 The case should be avoided for Al2 O3 microfiltration membrane supports.17 In the present work, the powders with tri-modal particles size distribution were prepared by mixing three different powder sizes (d50 = 37 ␮m, 8.2 ␮m, 1.6 ␮m). The effect of particle size distribution on the pore size distribution, the porosity and bending strength of ceramic membrane supports were investigated in detail. 2. Experimental 2.1. Preparation of ceramic membrane supports Al2 O3 powders, labeled as W50, W10 and W1 (purity of 99.5%, Henan White-dove Group, China) were purchased without further treatment prior to the preparation of microfiltration membrane supports. The d50 of W50, W10 and W1 were 37 ␮m, 8.2 ␮m and 1.6 ␮m, respectively. Their particle size distributions are shown in Fig. 1. Alumina powders with different average particle size were prepared by mixing W50, W10 and W1 in different ratios by ball milling for 2 h at 150 rpm as shown in Table 1. The particle size distribution and loose density of the mixed powders were measured. To prepare the support, alumina powder, 2 wt% TiO2 were mixed with 1 wt% poly vinyl alcohol (PVA, molecular weight

of 1750 ± 50, was solved into 5 wt% aqueous solution) of the total alumina powder in a ball mill at 150 rpm for 2 h to avoid the breakage of alumina grains. The mass ratio of powder:zirconia ball:alcohol is 1:1:2. The obtained suspensions were dried in an oven at 70 ◦ C for overnight. The mixed powders were compressed into the rectangle shape of 40 mm × 10 mm × 10 mm (L × h × w) by dry pressing under the applied pressure of 12 MPa. The samples were calcined in a programmable furnace at a heating rate of 3 ◦ C/min to 1650 ◦ C and annealing for 2 h. 2.2. Characterization The particle size distributions were determined by laser scattering particle size analyzer (Bettersize2000, Dan Dong, China). The loose densities of the powders were measured by Hall flow meter (FT-01, Ningbo, China). The densities of the green compacts were measured by mass-volume method. The dimensions were measured by a caliper and the weight was measured by the electronic balance with the accuracy of 0.0001 g. The pore size distributions and the porosity of the sintered compacts were measured by Mercury Intrusion porosimeter (Autopore IV9500, Micromeritics, USA). The bending strength was determined by three-point bending method at room temperature using universal material testing machine (WDW-30, Xi’an Letry Machine Testing Co. Ltd., China). The span length is 20 mm and the loading speed is 0.2 mm/min. Three samples were tested to determine the bending strength. Fracture surfaces of the sintered ceramic supports were observed by means of field emitting scanning electron microscope (JSM-6700F, JEOL, Japan). To test the evolution of the shape of alumina grain, the green compact of sample S4 was observed by SEM. 3. Results and discussion 3.1. Relationship of particles size distribution with pore size distribution of supports The particle size distributions of the mixed powders S1–S9 are shown in Fig. 2. All particle sizes distribution of the mixed powders shows the tri-modal distribution. The difference in mass

Please cite this article in press as: Chang Q, et al. Effect of particle size distribution of raw powders on pore size distribution and bending strength of Al2 O3 microfiltration membrane supports. J Eur Ceram Soc (2014), http://dx.doi.org/10.1016/j.jeurceramsoc.2014.06.001

+Model JECS-9695; No. of Pages 7

ARTICLE IN PRESS Q. Chang et al. / Journal of the European Ceramic Society xxx (2014) xxx–xxx

Fig. 2. Particles size distribution of S1–S9 powders.

3

percentages between W50, W10 and W1 is correspondingly reflected by the particle size distributions of the mixed powders. Among of them, the particle size distribution curves overlap in the range of 20–100 ␮m due to the unchanged mass percentage of W50. The corresponding relationship indicates the mixing processes of the powders have a negligible effect on the particle size distribution. To indicate the correspondence between the particle size distribution of the mixed powders and the pore size distribution of the supports, the labels of the supports are termed as S1-S9 supports in this manuscript, respectively. The pore size distributions of S1–S9 supports are shown in Fig. 3. It can be seen that there is less difference in pore size distributions of S1–S3 supports using S1–S3 powders as raw materials (labeled as S1 support for convenient, similarly hereafter) regardless of the ratio of W10/W1. The mean pore sizes of the S1–S3 support are all about 9.60 ␮m. It can be explained that W50 grains takes 80 wt% of the S1–S3 supports, which is absolutely dominant in the supports. W50 grains independently form the framework of the supports by a point-to-point particle packing, which determines the pore size distribution of the supports. Fine W10 and W1 grains only fill up the interstices among of coarse W50 grains. The interstices are large enough to accommodate all the W10 and W1 grains. After sintering, W1 grains maybe disappear according to the sintering theory.12–16 Therefore, the pore size distribution of the support is insensitive to the W10/W1 ratio. For S4–S6 supports, the mean pore sizes increase from 8.57 ␮m to 9.82 ␮m and 10.37 ␮m with the increase of W1 content. At the same time, the pore size distribution changes slightly wider. The mean pore size of S4 support is smaller than those of S1–S3 supports; however, S5 and S6 are the opposite. It may be explained that when W50 content is down to 70%, the interstices formed by W50 grains cannot accommodate W10 and W1 grains. W50 grains cannot independently form the framework of the supports. Some of W50 grains are connected through the bridge of W10 and W1 grains. The amount of W10 or W1 grains in the framework of the support depends on the volume and the amount of W10 grains and W1 grains. For S4 support, W10 grains are mainly involved in the framework. The sintering shrinkage of W10 grains leads to the shrinkage of the framework of S4 support and results in the decrease of the mean pore size. Lange21 had observed that the smaller grain would break-away from one of the larger grains and be absorbed by the other if the smaller grain was to redistribute its mass to the larger grains and the larger grains were constrained from moving together. For S5 and S6 support, W1 grains are involved the framework and may situate between two coarse grains. Thus, two pores would be united into a whole pore due to the break of the bonding W1 grains with W50 grains in framework. The pore size of support changes larger. The pore size distribution curves shift to the direction of large pore and the curves broaden at the same time. The similar phenomenon was observed by Wang et al.22 for membranes caused by the increase of sintering temperature. When the content of W10 and W1 grains increases further, more W10 and W1 grains are involved in the framework. The mean pore size of S7–S9 support changes irregularly from 8.65 ␮m to 9.82 ␮m and 9.37 ␮m. For S7 support, the mean pore

Please cite this article in press as: Chang Q, et al. Effect of particle size distribution of raw powders on pore size distribution and bending strength of Al2 O3 microfiltration membrane supports. J Eur Ceram Soc (2014), http://dx.doi.org/10.1016/j.jeurceramsoc.2014.06.001

+Model JECS-9695; No. of Pages 7 4

ARTICLE IN PRESS Q. Chang et al. / Journal of the European Ceramic Society xxx (2014) xxx–xxx

Fig. 4. SEM image of the green compact molded by S4 powder.

size depends on the combined action of the framework shrinkage caused by the sintering of W10 and the framework breakage caused by the sintering of W1, as discussed above. The framework breakage caused by the sintering of W1 is the dominant in S8 support. If the W1 content increases further, for S9 support, W1 grains can connect each other. The sintering among W1 grains is priority to the sintering between W50 and W1 grains.23 The sintering shrinkage of W1 grains themselves results in the deformation and the rearrangement of the whole framework. Therefore, the mean pore size of S9 support decreases. Fig. 4 shows the SEM image of the green compact molded using S4 powder. As can be seen, Most of W50 grains connect each other by point-by-point and form the framework. Some of W10 grains fill up the interstices (as shown by the left arrow in Fig. 4). Some W10 grains bridge coarse W50 grains (as shown by the right arrow in Fig. 4). Most of W1 grains distribute on the surface of W50 grains or W10 grains. In the image, there is no W1 agglomeration whose size is comparable to W10 or W50 grain. Therefore, the sintering of W1 grains obeys the sintering rule described in Ref.12–16 3.2. Relationship of particles size distribution with porosity of supports

Fig. 3. Pore size distribution of the ceramic membrane supports using S1–S9 powders as raw materials (sintering temperature is 1650 ◦ C, annealing time is 2 h).

The particle size distribution of raw material has an effect on the sintering and porosity of support by the powders packing.17,24,25 Powder packing can be characterized by loose density and green compact density. In order to be consistent with the porosity of supports, the porosity of loose powders and green compacts were calculated according to the loose density and the green compact density given the density of alumina is 3.9 g/cm3 . Fig. 5 shows the porosities of the loose powders, the green compact and the supports using S1–S9 powders as raw powders. It can be seen that the porosities of the loose powders and the supports increase with increasing of W1 content for the group of the samples with the constant mass percentage of W50, and the porosity of the green compact is the opposite. The porosity of the loose powder reflects the freedom

Please cite this article in press as: Chang Q, et al. Effect of particle size distribution of raw powders on pore size distribution and bending strength of Al2 O3 microfiltration membrane supports. J Eur Ceram Soc (2014), http://dx.doi.org/10.1016/j.jeurceramsoc.2014.06.001

+Model JECS-9695; No. of Pages 7

ARTICLE IN PRESS Q. Chang et al. / Journal of the European Ceramic Society xxx (2014) xxx–xxx

5

accumulation and the porosity of the green compact reflects the restrict accumulation of powders. The freedom accumulation can change into the restrict accumulation by pressing. The loose densities and the green compact of W50, W10 and W1 grain are 1.39 g/cm3 , 1.03 g/cm3 , 0.75 g/cm3 and 2.54 g/cm3 , 2.82 g/cm3 , 3.02 g/cm3 , respectively. Corresponding, the calculated porosity of loose powders and green compacts are 64.4%, 73.6%, 80.8% and 34.9%, 27.7%, 22.6%. The porosities of loose S1–S3 powders and S1–S3 green compacts decrease slightly comparing to those of W50 grain, indicating S1–S3 powders has the similar framework to that of W50 powder. It means that seldom fine grains are involved in the framework of S1–S3 compacts. However, for S4–S6 and S7–S9, the porosities of loose powders increase and the porosities of green compacts decrease. It can be explained that fine W10 and W1 grains are evolved in the framework. Fine W1 grains tend to form the agglomeration with high porosity. But W1 grains in the agglomeration are easily rearranged by pressures. Therefore, thus, the porosity of loose powder increases but porosity of green compacts decreases with the increasing of the W1 grain content. The porosity of the sintered compact depends on the distribution of W1 grains. As discussed above, W50 is dominant in S1–S3 support and W1 grains are very few. W1 grains discretely distribute on the surface of W50 grains. W1 grains may disappear during the sintering process and leave some interstices in the sintered compact. The porosities of S1–S3 increase slightly because the framework (W50 grains) has the little sintering shrink. However, the large sintering shrinkages are generated with the increase of W1 content increases in S4–S6 and S7–S9 support due to the participation of W1 aggregation in the framework of the support. The internal sintering of W1 grains aggregation leads to the deformation or the collapse of the framework of the supports. Therefore, the porosities of the supports depend on the distribution of W1 grains and the particles size distribution of the raw materials. 3.3. Relationship of particles size distribution with bending strength of supports As porous ceramic, the bending strength of microfiltration membrane supports depends on the bonding among the grains in support’s framework. Fig. 6 shows the bending strengths of S1–S9 supports as a function of mass ratio of W10/W1. The bending strength decreases with increasing the W1 content under the condition of a given W50 content. However, the regularity of the bending strength with the W50 content is not obvious for the supports. The mechanical properties of porous ceramic have been experimentally formularized as a function of porosity, which is quoted in.26 Accordingly, the empirical relationship is given by the following equation: σ = σ0 exp(−βε) Fig. 5. Porosities of loose powder, green compact and sintered compact using S1–S9 as raw powders.

(1)

where σ and σ 0 are the strengths of ceramics with and without pores, respectively, β is the parameter determined by the nature of porosity, and ε is the porosity. According to Eq. (1), the strength of porous ceramics decreases exponentially with

Please cite this article in press as: Chang Q, et al. Effect of particle size distribution of raw powders on pore size distribution and bending strength of Al2 O3 microfiltration membrane supports. J Eur Ceram Soc (2014), http://dx.doi.org/10.1016/j.jeurceramsoc.2014.06.001

+Model JECS-9695; No. of Pages 7 6

ARTICLE IN PRESS Q. Chang et al. / Journal of the European Ceramic Society xxx (2014) xxx–xxx

Fig. 6. Bending strengths of the supports prepared using S1–S9 as raw powders.

porosity. Correspondingly, it can be deduced that porosity has little effect on the bending strength if porosity is over 30%. Under this condition, the pores can be regarded as a continuous phase. And then, the porous ceramic is regarded as a three-dimensional network connected by grains in the framework. The bending strength of porous ceramics depends on the neck area between grains and the amount of the necks. S1–S3 supports have the lower bending strength because of the poor sintering of coarse W50 grains. Fig. 7(A) shows the SEM image of S1 support. As can be seen, W50 grains form the framework of S1 support by a point-by-point connection. The contact interface of the neck is flat as shown in Fig. 7. Obviously, the connecting areas among the grains are far smaller than the outer surface of W50 grains. Simultaneously, the amount of sintering necks is not enough based on the fact that the porosity is the highest among the supports with different content of W50. Therefore, the bending strength of S1 support is far lower than that of S4 or S7 support. S4 support has the highest bending strength than the other supports. Fig. 7(B) shows the SEM image of S4 support. As can be seen, W10 grains are combined firmly to the neighboring W50 grains through the sintering necks, which is different from S1 support. The enhanced neck growth between W10 grains and W50 grains increases the bending strength of the support sintered at a relatively lower sintering temperature. However, if W1 grains take the place of W10 grains, the bending strength decreases sharply. The reason is that the fine grains will disappear, which weakens or breaks the bonding among W50 grains.19 The framework is still kept because the W1 content in the framework is small. Therefore, the porosity of the sintered compact changes slightly but the bending strength decreases sharply. For S7–S9 supports, the bending strengths are larger than those of S1–S3, S5 and S6 supports due to the decrease of porosity according to Eq. (1). Fig. 7(C) shows the SEM image of S7 support. It can be seen that W50 grains are bridged by W10 and W1 grains. After sintering, aggregated W1 grains coarsen and contact with W50 grains by the points or surface contact. Here, the amount of the sintering necks is large enough because

Fig. 7. SEM images of (A) S1 support, (B) S4 support and (C) S7 support.

W1 grains involve in the framework. Similarly, the sintering shrinkage of W1 grains decreases the porosity of the supports, as discussed above. Therefore, the bending strength of S7 or S8 support is higher than that of S1 or S2 support. However, the bending strength of S9 support is lower than that of S3 support even if the W1 content increases. It can be explained that the sintering among fine W1 grains is priority to the sintering of W10 or W50 grains. W1 grains do not promote the sintering of W10

Please cite this article in press as: Chang Q, et al. Effect of particle size distribution of raw powders on pore size distribution and bending strength of Al2 O3 microfiltration membrane supports. J Eur Ceram Soc (2014), http://dx.doi.org/10.1016/j.jeurceramsoc.2014.06.001

+Model JECS-9695; No. of Pages 7

ARTICLE IN PRESS Q. Chang et al. / Journal of the European Ceramic Society xxx (2014) xxx–xxx

or W50 grains as a sintering-aid. The neck area in the framework is not large though the amount of the sintering necks increases with the W1 grains content, as shown in Fig. 7(C). The support may be crushed gradually through breaking the weak bonding of W1 grains. 4. Conclusions To lower the sintering temperature of Al2 O3 microfiltration membrane support, Al2 O3 powders with tri-modal of particle size distribution are used. The particle size distribution of Al2 O3 powders has an important effect on the pore size distribution and bending strength of Al2 O3 membrane supports. If the fine grains distribute discretely, the bending strength of supports increases with the fine grains content along with the slight increase in porosity. The aggregated fine grains are harmful to the bending strength and the pore size distribution of supports. Medium grains have a positive effect on the bending strength through a bridging function between coarse grains, but have less effect on porosity. The addition of medium/fine grains contributes to increase the bending strength but do not make the pore size distribution wide if only the coarse grains forms the support’s framework. The bending strength of S4 support (70 wt% W50, 24 wt% W10 and 6 wt% W1) sintered at 1650 ◦ C is 49.8 MPa and the pore size distribution is very narrow. To obtain the similar bending strength, 1720 ◦ C is required for 100 wt% W50. It indicates the appreciate particle size distribution of the raw powders contribute to lower the sintering temperature and has less effect on the pore size distribution. Acknowledgments The authors gratefully acknowledge the financial support provided by Sino-French International Science and Technology Cooperation Program (No. 2011DFA52000) and National Natural Science Foundation of China (Nos. 51262012, 51362015,51262012,51062006). References 1. Bae D, Cheong D, Han K, Choi S. Fabrication and microstructure of Al2 O3 /TiO2 composite membranes with ultrafine pores. Ceram Int 1998;24:25–30. 2. Li S, Wang CA, Zhou J. Effect of starch addition on microstructure and properties of highly porous alumina ceramics. Ceram Int 2013;39:8833–9. 3. Dong YC, Lin B, Zhou JE, Zhang XZ, Ling YH, Liu XQ, et al. Corrosion resistance characterization of porous alumina membrane supports. Mater Charact 2011;62:409–18. 4. Monash P, Pugazhenthi G. Effect of TiO2 addition on the fabrication of ceramic membrane supports: a study on the separation of oil droplets and bovine serum albumin (BSA) from its solution. Desalination 2011;279:104–14.

7

5. Avci GG, Misirli Z, Gunay V. Processing and characterization of microfiltration supports prepared from alumina powders. Ceram Int 1996;22: 23–6. 6. Bouzerara F, Harabi A, Achour S, Larbot A. Porous ceramic supports for membranes prepared from kaolin and doloma mixtures. J Eur Ceram Soc 2006;26:1663–71. 7. Wang YH, Zhang Y, Liu XQ, Meng GY. Microstructure control of ceramic support from corundum–rutile powder mixture. Powder Technol 2006;168:125–33. 8. Vasanth D, Pugazhenthi G, Uppaluri R. Fabrication and properties of low cost ceramic microfiltration membranes for separation of oil and bacteria from its solution. J Membr Sci 2011;379:154–63. 9. Ananthakumar S, Menon ARR, Prabhakaran K, Warrier KGK. Rheology and packing characteristics of alumina extrusion using boehmite gel as a binder. Ceram Int 2001;27:231–7. 10. Chen GL, Qi H, Xing WH, Xu NP. Direct preparation of macroporous mullite support for membranes by in situ reaction sintering. J Membr Sci 2008;318:38–44. 11. Sarkar S, Banyopadhyay S, Larbot A, Cerneaux S. New clay-alumina porous capillary supports for filtration application. J Membr Sci 2012;392–393:130–6. 12. Darcovich K, Be’ra L, Shinagaw K. Particle size distribution effects in an FEM model of sintering porous ceramics. Mater Sci Eng A 2003;341:247–55. 13. Ch’ng HN, Pan JZ. Sintering of particles of different sizes. Acta Mater 2007;55:813–24. 14. Wakai F, Yoshida M, Shinoda Y, Akatsu T. Coarsening and grain growth in sintering of two particles of different sizes. Acta Mater 2005;53:1361–71. 15. Pan J, Le H, Kucherenko S, Yeomans JA. A model for the sintering of spherical particles of different sizes by solid state diffusion. Acta Mater 1998;46(13):4671–90. 16. Darcovich K, Toll F, Hontanx P, Roux V, Shinagawa K. An experimental and numerical study of particle size distribution effects on the sintering of porous ceramics. Mater Sci Eng A 2003;348:76–83. 17. Ma J, Lim LC. Effect of particle size distribution on sintering of agglomerate-free submicron alumina powder compacts. J Eur Ceram Soc 2002;22:2197–208. 18. Lu GQ. Evolution of the pore structure of a ceramic powder compact during sintering. J Mater Process Technol 1996;59:297–302. 19. Parhami F, McMeeking RM, Cocks ACF, Suo Z. A model for the sintering and coarsening of rows of spherical particles. Mech Mater 1999;31:43–61. 20. Shiau FS, Fang TT, Leu TS. Effect of particle size distribution on the microstructural evolution in the intermediate stage of sintering. J Am Ceram Soc 1997;80:286–90. 21. Lange FF. Densification of powder compacts: an unfinished story. J Eur Ceram Soc 2008;28:1509–16. 22. Wang P, Huang P, Xu NP, Shi J, Lin YS. Effect of sintering on properties of alumina microfiltration membranes. J Membr Sci 1999;155:309–14. 23. Lee HM, Huang CY, Wang CJ. Forming and sintering behaviors of commercial ␣-Al2 O3 powders with different particle size distribution and agglomeration. J Mater Process Technol 2009;209:714–22. 24. Biesheuvel PM, Verweij H. Design of ceramic membrane support: permeability, tensile strength and stress. J Membr Sci 1999;156:141–52. 25. Azar M, Palmero P, Lombardi M, Garnier V, Montanaro L, Fantozzi G, et al. Effect of initial particle packing on the sintering of nanostructured transition alumina. J Eur Ceram Soc 2008;28:1121–8. 26. Hirata Y, Shimonosono T, Sameshima T, Sameshima S. Compressive mechanical properties of porous alumina powder compacts. Ceram Int 2014;40:2315–22.

Please cite this article in press as: Chang Q, et al. Effect of particle size distribution of raw powders on pore size distribution and bending strength of Al2 O3 microfiltration membrane supports. J Eur Ceram Soc (2014), http://dx.doi.org/10.1016/j.jeurceramsoc.2014.06.001